Sometimes, but rarely, when I post an answer that turns out to be incorrect and it makes more sense to just start over, I delete it and re-post a second. Of course, the first is still there, although it's deleted.

Other than something like that, though, are there acceptable reasons for adding more than one answer to a question?

  • I'd do this If i was downvoted for being wrong and fixing it would take long enough that it'd be downvoted to oblivion – Cole Johnson Aug 15 '13 at 23:09
  • @Cole but once you realized what was wrong, you could delete the question, edit it, then undelete it. Posting a second, different answer (even if you deleted the first) leaves a bunch of extra baggage on the question that 10K/mods have to look at every time. – Aaron Bertrand Aug 15 '13 at 23:11
  • 1
  • @AaronBertrand I thought we couldn't edit our deleted posts? – Cole Johnson Aug 15 '13 at 23:14
  • @ColeJohnson I can definitely edit deleted questions. I just tried it now on SO. – Jason C Aug 15 '13 at 23:15
  • @Cole did you try it? – Aaron Bertrand Aug 15 '13 at 23:17
  • @AaronBertrand no, because I've never been in a situation like what I described – Cole Johnson Aug 15 '13 at 23:20
  • @ColeJohnson Ha. It's tough never being wrong, isn't it? :) – Jason C Aug 15 '13 at 23:27
  • @JasonC I know. I still dream of the day I become Jon Skeet ;) – Cole Johnson Aug 15 '13 at 23:56
  • @ColeJohnson data.stackexchange.com/stackoverflow/query/3160/… – Jason C Aug 16 '13 at 0:01
  • 1
    @JasonC I got a 1 and Skeet got a 6 :( – Cole Johnson Aug 16 '13 at 0:26

I do this with some regularity. If each answer provides a distinct method it is arguable that separating them allows the community to vote for the one that is better rather than voting for an omnibus answer.

Of course this could easily be overused and most of the time I post a single answer with multiple options. I try to reserve posting two answers for cases where the approach is starkly different.

| improve this answer | |
  • 1
    And now I notice that this question is apparently a duplicate. :-/ – Mr.Wizard Aug 15 '13 at 23:08
  • Hmm, I'm not sure about this. I'd say that questions where a person can post two completely separate answers with different solutions, but each answer still be correct and useful, are probably not acceptable questions for Stack Exchange sites. That is, they are probably polls, or have underspecified requirements. – Cody Gray Aug 15 '13 at 23:11
  • @Cody I don't agree with that. In SQL Server, for example, there are many ways to do something, and some of them work better on (or require) a specific version, some work better on (or require) a different edition, etc. It's not something I ever do but I don't think this scenario automatically makes the question unacceptable. – Aaron Bertrand Aug 15 '13 at 23:13
  • @CodyGray As respectfully as is possible for this to be: I have no idea what you're talking about. Beta sites are encouraged to have more than 2.5 answers per question. Many if not most questions can have more than one reasonable answer. Or do I misunderstand you? – Mr.Wizard Aug 15 '13 at 23:13
  • I have no idea why that is. The questions we're looking for tend to have one, possibly two, answer at most. – Cody Gray Aug 15 '13 at 23:14
  • 1
    @Mr.Wizard right, if two people can post two completely separate answers, and both are valid ways to do something, why can't one person do that? – Aaron Bertrand Aug 15 '13 at 23:14
  • @CodyGray I'm sorry but you have an impoverished perspective on this. Come over to Mathematica some time and you'll see that it is quite possible for a question to have half a dozen good answers. (If course that is the exception but I stand by my claim.) – Mr.Wizard Aug 15 '13 at 23:16
  • 2
    @Cody I think that's a very bizarre limitation and not a very well-publicized goal if it is in fact part of the mission of this site. While the goal for one specific OP is to have a "best" answer, there is no requirement that they choose one, and different answers can be better for different future readers in different circumstances. – Aaron Bertrand Aug 15 '13 at 23:16
  • I do think I see where Cody is coming from, which relates to a comment I just made on John's answer here. SO tends to reject questions that don't have concrete answers (not because they're bad, but because I guess that's what SO goes for). If a question has two completely different answers that might be an indication that the question was too broad. When I have two (or more) different answers I usually post them as one and outline the differences. I can see how two answers might make separating upvotes easier, but you lose some ability to write out a comparison. – Jason C Aug 15 '13 at 23:20
  • I don't agree that the fact that a question has multiple potential answers means it is too broad. Can you really say, with a straight face, that every programming problem can only be solved in one way? If that were the case, maybe we should consider only allowing one answer per question, period? – Aaron Bertrand Aug 15 '13 at 23:47
  • @Mr.Wizard no, I'm asserting that it is a fact that there are valid questions with multiple potential solutions (and sometimes the same person might even dream up these multiple solutions on their own). I don't agree that the fact that such questions exist "might be an indication that the question was too broad." Not sure if that my line of thinking any clearer, but it's the best I can do. :-) – Aaron Bertrand Aug 16 '13 at 0:18
  • @AaronBertrand I can say (with a straight face) that almost every programming problem can be solved with a bit of analysis, sane judgment, and some common sense. The vast majority of questions that I see on SO relate to the asker having difficulty, in one form or another, translating the concept they have in their head to whatever language they are programming in. There may be very different specific solutions, but generally the same thought process is required to arrive at each, and that process is the real right answer. But, I'm 100% aware of how unpopular my "unified theory" of things is. – Jason C Aug 16 '13 at 2:00
  • @JasonC but are you implying that two different people can't come to different answers even if they've both employed analysis, sane judgment, and some common sense? I don't always agree with the experts in my field, and we often solve problems in different ways, but that doesn't make one of us more right than the other. Many ways to skin a cat, and all that... – Aaron Bertrand Aug 16 '13 at 2:05
  • 1
    I am not a cat person, by the way. I just don't know enough about skinning them to make my example more concrete. – Aaron Bertrand Aug 16 '13 at 2:19
  • 1
    @JasonC Not saying it's a bad idea to offer multiple ideas in a single answer, but it's not the norm, should not be expected, and when I do it (yes, I have done it) I usually do so in an effort to show why one approach is better than the others, usually from a performance perspective, or what the user can do with their current limitations, and what they will be able to do later when those limitations are lifted. – Aaron Bertrand Aug 16 '13 at 2:34

I think it's acceptable to add more than one answer to a question if ... you have more than one answer to a question.

If they're both helpful answers, why not? This doesn't seem any different than answering a question once, describing two completely different methods. That, and the SO engine allows it. If it were undesirable to do this, it would be disallowed, right?

Or, if you answer a question, and then some time lateryou happen upon a better way to do it, then it would be reasonable to add another answer. It would make interpretation of the upvotes less confusing (if anyone cares) because it would be clear that the upvotes were for the original answer, and not the original answer PLUS the answer you added later.

| improve this answer | |
  • Good point about separating the upvotes. Usually when I have two separate answers I post them as one answer and try to point out the differences between them. I suppose, though, if they were so different that a comparison makes no sense, it would make sense to post two answers -- although that case may also indicate that the original question was too broad to remain on SO. – Jason C Aug 15 '13 at 23:17
  • @JasonC as I commented above, I don't understand why that logic makes sense. If two independent people can post separate, different, but good answers, why couldn't one person do the same? – Aaron Bertrand Aug 15 '13 at 23:18
  • @AaronBertrand Yeah, it's a good point, and it's very gray. I feel like I haven't seen any questions on SO where there are two distinctly different good answers and the question is "appropriate" for SO -- but I also haven't been looking at answers that way (until now). Just to be clear though, I'm not pushing to disallow multiple answers or anything, just wondering about etiquette. I do feel strongly, though, that one person with two different answers could provide a much better answer by combining their two answers into one and highlighting the differences. :) – Jason C Aug 15 '13 at 23:24
  • 1
    That is usually what I do, but I think it's just as valid to separate independent answers. – Aaron Bertrand Aug 15 '13 at 23:43

Not the answer you're looking for? Browse other questions tagged .